计算机科学

计算机科学 ›› 2024, Vol. 51 ›› Issue (8): 160-167.doi: 10.11896/jsjkx.230500171

• 计算机图形学&多媒体 • 上一篇    下一篇

任务感知的多尺度小样本SAR图像分类方法

张睿, 王梓祺, 李阳, 王家宝, 陈瑶   

  1. 陆军工程大学指挥控制工程学院 南京 210007
  • 收稿日期:2023-05-25 修回日期:2023-09-13 出版日期:2024-08-15 发布日期:2024-08-13
  • 通讯作者: 李阳(solarleeon@outlook.com)
  • 作者简介:(3959966@qq.com)
  • 基金资助:
    江苏省自然科学基金(BK20200581)

Task-aware Few-shot SAR Image Classification Method Based on Multi-scale Attention Mechanism

ZHANG Rui, WANG Ziqi, LI Yang, WANG Jiabao, CHEN Yao   

  1. Command and Control Engineering College,Army Engineering University of PLA,Nanjing 210007,China
  • Received:2023-05-25 Revised:2023-09-13 Online:2024-08-15 Published:2024-08-13
  • About author:ZHANG Rui,born in 1977,Ph.D,professor,Ph.D supervisor.His main research interests include data enginee-ring and information fusion.
    LI Yang,born in 1984,Ph.D,associate professor,is a senior member of CCF(No.D24215).His main research in-terests include computer vision,deep learning and image processing.
  • Supported by:
    Natural Science Foundation of Jiangsu Province,China(BK20200581).

PDF (PC)

419

摘要： 针对SAR图像分类时存在的带标注样本较少的问题,提出了一种任务感知的多尺度小样本SAR图像分类方法。为了能够充分挖掘局部特征并关注具体任务下的关键局部语义patches,引入了两种有效的注意力机制,获得了更加高效且丰富的特征表示。首先,在特征提取阶段使用互补注意力模块(CSE Block),关注原始特征中不同语义部分的显著特征,从被抑制的特征中提取次级显著特征并与主要显著特征融合,得到更加高效且丰富的特征表示。随后,利用自适应情景注意力模块(AEA Block)获得整个任务中的关键语义patches,增强任务间的区分信息,提升小样本SAR图像分类任务的精度。结果表明,在SAR图像分类标准数据集MSTAR上,5-way 1-shot任务分类精度相较于次优方法精度提升了2.9%,并且该方法在两项任务中的运行时间与其他度量学习方法相比水平相当,未额外增加过多的计算资源,验证了其有效性。

关键词: 多尺度注意力机制, 小样本学习, SAR图像分类, 度量学习

Abstract: Aiming at the problem of the lack of labeled samples in SAR image classification,this paper proposes a task-aware few-shot SAR image classification method based on multi-scale attention mechanism.In order to fully mine local features and focus on the key local semantic patches under specific tasks,this paper introduces two effective attention mechanisms to obtain more efficient and rich feature representation.First,in the feature extraction stage,the complemented squeeze-and-excitation attention block(CSE Block) is used to focus on the salient features of different semantic parts of the original features.It can extract secon-dary salient features from the suppressed features and merge them with the main salient features,which can obtain more efficient and rich feature representation.Subsequently,an adaptive episodic attention block(AEA Block) is used to obtain key semantic patches in the entire task,which can enhance the differentiated information between tasks and improve the accuracy of SAR image classification tasks.The results show that the classification accuracy of the 5-way 1-shot task is 2.9% higher than that of the sub-optimal task on the SAR image classification standard MSTAR dataset.In the two tasks,the runtime of the proposed method is the same as other metric-learning methods,without additional excessive computing resources,which verifies its effectiveness.

Key words: Multi-scale attention mechanism, Few-shot learning, SAR image classification, Metric learning

中图分类号: 

  • TP391
[1]LI Y,WANG J B,XU Y L,et al.DeepSAR-Net:Deep convolutional neural networks for SAR target recognition[C]//2017 IEEE 2nd International Conference on Big Data Analysis(ICBDA).2017:740-743.
[2]GUO W W,ZHANG Z H,YU W X,et al.Perspective on explainable SAR target recognition[J].Journal of Radars,2020,9(3):462-476.
[3]SU S H,CUI Z T,GUO W W,et al.Explainable analysis of deep learning methods for SAR image classification[C]//IEEE International Geoscience and Remote Sensing Symposium(IGARSS).2022.
[4]GE Y Z,LIU H,WANG Y,et al.Survey on deep learning image recognition in dilemma of small samples[J].Journal of Software,2022,33(1):193-210.
[5]LIU Y,LEI Y B,FAN J L,et al.Survey on image classification technology based on small sample learning[J].Acta Automatica Sinica,2021,47(2):297-315.
[6]LIU D Y,GAO X Z,SHEN Q M.Prototypical network for radar image recognition with few samples[C]//Journal of Physics:Conference Series.IOP Publishing,2020.
[7]SUN X,LV Y X,WANG Z R,et al.SCAN:Scattering characteristics analysis network for few-shot aircraft classification in high-resolution SAR images[J].IEEE Transactions on Geo-science and Remote Sensing,2022,60:1-17.
[8]ZHANG L B,LENG X G,FENG S J,et al.Domain knowledge powered two-stream deep network for few-shot SAR vehicle recognition[J].IEEE Transactions on Geoscience and Remote Sensing,2021,60:1-15.
[9]DONG C Q,LI W B,HUO J,et al.Learning task-aware local representations for few-shot learning[C]//Proceedings of the Twenty-Ninth International Conference on International Joint Conferences on Artificial Intelligence.2021:716-722.
[10]LI F F,FERGUS R,PERONA P.A Bayesian approach to unsupervised one-shot learning of object categories[C]//Proceedings of the IEEE International Conference on Computer Vision.2003:1134-1141.
[11]PENG Y C,QIN X L,ZHANG L G,et al.Survey on Few-shot Learning Algorithms for Image Classification[J].Computer Science,2022,49(5):1-9.
[12]ZHAO K L,JIN X L,WANG Y Z.Survey on few-shot learning[J].Journal of Software,2021,32(2):349-369.
[13]ZHANG R,YANG Y X,LI Y,et al.Multi-task few-shotlearning with composed data augmentation for image classification[J].IET Computer Vision,2022,17(2) 211-221.
[14]LIU X,ZHOU K R,HE Y L,et al.Survey of metric-based few-shot classification[J].Pattern Recognition and Artificial Intelligence,2021,34(10):909-923.
[15]SUNG F,YANG Y X,ZHANG L,et al.Learning to compare:Relation network for few-shot learning[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2018:1199-1208.
[16]LI W B,WANG L,XU J L,et al.Revisiting local descriptorbased image-to-class measure for few-shot learning[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2019:7260-7268.
[17]ZHAO X,WANG J B,LI Y,et al.Complemented attentionmethod for fine-grained image classification[J].Journal of Image and Graphics,2021,26(12):2860-2869.
[18]ZHAO P F,HUANG L J,XIN Y,et al.Multi-aspect SAR target recognition based on prototypical network with a small number of training samples[J].Sensors,2021,21(13):4333.
[19]LECUN Y,BENGIO Y,HINTON G.Deep learning[J].Nature,2015,521(7553):436-444.
[20]AL-ANTARY M T,ARAFA Y.Multi-scale attention network for diabetic retinopathy classification[J].IEEE Access,2021,9:54190-54200.
[21]FAN T L,WANG G L,LI Y,et al.Ma-net:A multi-scale attention network for liver and tumor segmentation[J].IEEE Access,2020,8:179656-179665.
[22]ZHANG J Y,CHEN X S,QIU Z X,et al.Hard exudate segmentation supplemented by super-resolution with multi-scale attention fusion module[C]//2022 IEEE International Conference on Bioinformatics and Biomedicine(BIBM).2022:1375-1380.
[23]CHEN H Z,LI Y N,FANG H J,et al.Multi-scale attention 3D convolutional network for multimodal gesture recognition[J].Sensors,2022,22(6):2405.
[24]WANG Y Q,YAO Q M,KWOK J T,et al.Generalizing from a few examples:A survey on few-shot learning[J].ACM Computing Surveys,2020,53(3):1-34.
[25]JIAN Y R,TORRESANI L.Label hallucination for few-shotclassification[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2022:7005-7014.
[26]HU J,SHEN L,SUN G.Squeeze-and-excitation networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2018:7132-7141.
[27]WANG H P,CHEN S Z,XU F,et al.Application of deep-lear-ning algorithms to MSTAR data[C]//2015 IEEE International Geoscience and Remote Sensing Symposium(IGARSS).2015:3743-3745.
[28]KINGMA D P,BA J.Adam:A method for stochastic optimization[C]//International Conference on Learning Representations(ICLR).2015.
[29]CHEN W Y,LIU Y C,KIRA Z,et al.A closer look at few-shot classification[C]//International Conference on Learning Representations(ICLR).2019.
[30]RAJASEGARAN J,KHAN S,HAYAT M,et al.Self-super-vised knowledge distillation for few-shot learning[C]//British Machine Vision Conference(BMVC).2021.
[31]FINN C,ABBEEL P,LEVINE S.Model-agnostic meta-learning for fast adaptation of deep networks[C]//International Confe-rence on Machine Learning.2017:1126-1135.
[32]GORDON J,BRONSKILL J,BAUER M,et al.Meta-learning probabilistic inference for prediction[C]//International Confe-rence on Learning Representations(ICLR).2019.
[33]BERTINETTO L,HENRIQUES J F,TORR P H S,et al.Meta-learning with differentiable closed-form solvers[C]//International Conference on Learning Representations(ICLR).2019.
[34]SUN Q,LIU Y Y,CHUA T S,et al.Meta-transfer learning for few-shot learning[C]//Proceedings of the IEEE/CVF Confe-rence on Computer Vision and Pattern Recognition.2019:403-412.
[35]RUSU A A,RAO D,SYGNOWSKI J,et al.Meta-learning with latent embedding optimization[C]//International Conference on Learning Representations(ICLR).2019.
[36]RAGHU A,RAGHU M,BENGIO S,et al.Rapid learning or feature reuse? Towards understanding the effectiveness of maml[C]//International Conference on Learning Representations(ICLR).2020.
[37]OH J,YOO H,KIM C H,et al.BOIL:Towards representation change for few-shot learning[C]//International Conference on Learning Representations(ICLR).2021.
[38]SNELL J,SWERSKY K,ZEMEL R.Prototypical networks for few-shot learning[J].arXiv:1703.05175,2017.
[39]YE H J,HU H,ZHAN D C,et al.Few-shot learning via embedding adaptation with set-to-set functions[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2020:8808-8817.
[40]LI W B,XU J L,HUO J,et al.Distribution consistency based covariance metric networks for few-shot learning[C]//Procee-dings of the AAAI Conference on Artificial Intelligence.2019:8642-8649.
[41]LIU Q,ZHANG X Y,LIU Y X.Few-shot SAR target recognition method based on gated multi-scale matching network[J].Systems Engineering and Electronics,2022,44(11):3346-3356.
[42]LI W B,WANG Z Y,YANG X S,et al.LibFewShot:A comprehensive library for few-shot learning[J].arXiv:2109.04898,2021.
[43]ZHANG H Y,ZHANG J,HUANG J.Multi-label Image Classification Model Based on Graph Attention Network[J].Journal of Chongqing Technology and Business University(Natural Science Edition),2022,39(1):34-41.
[44]LI X X,AN W J,WU J J,et al.Channel attention bilinear metric network[J].Journal of Jilin University(Engineering and Technology Edition),2024,54(2):524-532.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发